Who signs an electrical certificate, and what makes one worthless

The signature is the point of the document. Everything above it is findings; the bit at the bottom is somebody putting their name to those findings and accepting that they stand behind them.

Which makes it worth knowing exactly what you're signing, how many times, and what the law has to say about who's allowed to. That last one has a short answer: almost nothing.

The boxes, and what each one commits you to

An Electrical Installation Certificate carries three declarations. For design, for construction, and for inspection and testing. There are four signature lines, because the design block has room for a second designer where responsibility is shared.

One person can sign all three. The notes on the form say so plainly: a signatory authorised to certify more than one category of work signs in each of the appropriate places. So a sole trader who designed it, installed it and tested it signs three times. On a bigger job those are three different people, once each.

Each declaration carries the same limiting clause: the extent of liability of the signatory is limited to the work described as the subject of the certificate. You are not putting your name to the rest of the installation, and that is deliberate.

A Minor Works certificate has one signature, covering design, construction, inspection and testing together. It is for work that doesn't include a new circuit, and explicitly not for a consumer unit or distribution board change.

An EICR has two. One for the person who inspected and tested, and a separate one authorising the report for issue. The schedule of test results then carries a third, for whoever did the testing. And the reason a report has two while a certificate has three is that they are different kinds of document: a certificate asserts that work complies, while a report gives an opinion bounded by what was actually looked at.

One more difference people trip over. On an EIC the next-inspection field is an interval, recommended by the designer. On an EICR it is a date, and it has to be supported by a reason.

What the law says about who signs: almost nothing

This surprises people, so it's worth being precise.

Search the Building Regulations for "signature" and you get nothing. Search for "signed" and every hit attaches to the client or applicant — building notices, full plans, completion notices — not to the certificate an electrician issues.

The self-certification machinery is framed around a registered business, not a qualified individual. The regulation authorises building control to accept a certificate by "the person carrying out the work", requires it within 30 days, and says it is evidence, but not conclusive evidence, of compliance. Schedule 3 lists the schemes, and what it lists is corporate registration.

Even record-keeping isn't a legal duty in itself. HSE's own guidance on the Electricity at Work Regulations says it directly: records can help demonstrate compliance, although keeping records is not a legal requirement.

What the Electricity at Work Regulations do say is about the work, not the paperwork: nobody is to be engaged in work where technical knowledge or experience is needed to prevent danger unless they have it — or are under an appropriate degree of supervision. Knowledge or experience. And supervision is an alternative to both.

There are two places where the law gets closer.

Wales. Its guidance, unlike England's, actually uses the word: the certificate must be made out and signed only by someone qualified to do so — and then defines qualified as having the appropriate qualifications, knowledge and experience to carry out the testing and complete the certificate. Competence, not registration.

Scotland, for rented housing. The Housing (Scotland) Act requires the inspection to be carried out by a competent person, and requires the record to include the name, address and relevant qualifications of the person who carried it out. That is the closest any UK statute comes to regulating who signs.

The rented-sector rules elsewhere are circular by design. England's regulations define a "qualified person" as a person competent to do the work. Scotland's statutory guidance is the clearest of the lot: landlords may take membership of SELECT, NICEIC or NAPIT as evidence, but a skilled person outside those schemes should equally be able to confirm the checklist.

What BS 7671 asks for

Not a qualification. A skilled person (electrically): someone with adequate education, training and practical skills for the work, able to perceive risks and avoid the hazards electricity creates.

Worth knowing, because people still quote the old wording: BS 7671 used to define a "competent person" and dropped that definition in 2015, replacing it with skilled and instructed persons. The industry assessment specification that sits behind the schemes sets the minimum competence of an inspector as exactly that — satisfying the definition of a skilled person.

If you want the wider point about the standard's legal status, the EICR codes guide covers why BS 7671 isn't law and why that doesn't help you much.

What the schemes add

This is where most of what people believe to be rules actually lives.

Schemes register businesses, not people. NICEIC says so itself: it certifies businesses, not individuals, unless they are a sole trader, and responsibility for the work sits with the business. Registration is per contracting office, and it isn't transferable.

The qualified supervisor is a scheme construct. The assessment specification defines one as a competent person with day-to-day responsibility for the safety, technical standard and quality of the work, and requires at least one per office — and it allows one individual to hold that role alongside being the principal duty holder, which is how a one-person business satisfies it. NAPIT uses a different name for the same idea: technically competent individual.

And the supervisor does not have to have been there. NICEIC's own guidance describes certificates being signed by the individuals responsible for design, construction, inspection and testing, then reviewed and countersigned by the business's qualified supervisor, who is ultimately responsible — and says the supervisor does not need to physically inspect the work in order to sign their section.

That is a perfectly ordinary arrangement, and it's worth separating from the thing that isn't. Certsure's scheme rules say that unless you carried out the work and it was in scope of your certification, you don't issue their forms at all.

Apprentices. No scheme says an apprentice may not sign. What they set is a competence floor the apprentice hasn't reached yet: apprentices, mates and improvers sit at the instructed-person level, while the inspector has to be a skilled person. An apprentice can install under supervision, and can test under close supervision to build experience. They are not the inspector.

Scotland works differently again. Where work needs a building warrant, certificates of construction are signed by an approved certifier of construction and countersigned by the certification co-ordinator — two signatures, and a separate registration from ordinary scheme membership. Only two bodies run that scheme for electrical installations: SELECT and NICEIC. Whether a warrant is needed at all is a separate question, and usually the answer is no.

What actually makes a certificate worthless

Rarely the signature. Usually one of these.

No schedules. The EIC states that the schedules listed are part of the document and that the certificate is valid only when they are attached — and that it is valid only if the schedule of inspections has been completed and the schedules of circuit details and test results attached. The EICR says the same. A single sheet with nothing behind it isn't a tidy certificate; it's an incomplete one.

An extent-and-limitations box that swallows the job. On an EICR this is section D, and it is the part most worth reading. The form requires the extent and any limitations to be recorded with reasons, agreed with a named person, and agreed before the inspection rather than invented afterwards — there's a field on the form for who agreed them. The industry best-practice guide puts the consequence better than I could: the greater the limitations applied, the lesser the scope of the inspection, and the value of the report is correspondingly diminished.

Failing the whole report on the supplier's kit. The inspection schedule says that an outcome against the distributor's intake equipment, other than access to live parts, shouldn't be used to determine the overall outcome. It happens anyway.

A summary of condition that says nothing. "Good", or "needs a new fuse box", isn't a description of an installation's condition. The IET has been saying so for years, and the industry best-practice guide treats minimal descriptions as unacceptable, on the ground that they don't indicate the true condition of the installation.

The original going to the wrong place. The original goes to the person ordering the work; you keep the duplicate. How long to keep it is its own question.

And one that isn't a defect at all, despite what you'll be told:

Calibration. BS 7671 requires test instruments to be selected to the relevant parts of BS EN 61557. It does not require calibration — the word appears nowhere on any of the model forms. What the schedule of test results asks for is instrument serial or asset numbers. Nor is calibration what HSE's GS38 is about; that guidance is about probes, leads and fuses.

None of which means don't calibrate. The industry guidance is explicit that you should have an effective system to confirm and record continuing accuracy, and formal calibration at the manufacturer's recommended interval is the usual way to do it — and your scheme may require it outright. Just know which of those is the standard talking and which is your scheme.

Two details worth having. The assessment specification behind the schemes makes calibration mandatory in exactly one situation — instruments you hire — and otherwise asks for a system that demonstrates continuing accuracy, which can be cross-checks between your own instruments or measurements against a reference circuit. And the twelve-month interval everyone quotes isn't in any law, standard or scheme rule I could find. When a UKAS-accredited lab prints a calibration due date, it does so as the customer's stated request rather than as its own judgement.

Does it have to be on paper?

No rule I could find says either way, in either direction.

The model forms are silent — no mention of electronic signatures, of email, or of what makes something the "original". The law prescribes no format either: the regulations simply say give the occupier a copy. General law treats an electronic signature as capable of the same effect as a handwritten one, and both major schemes run their own electronic certification systems that will send a customer an electronic copy.

So a PDF isn't second-class. What I can't point you at is a clause in BS 7671 that blesses it, because the standard is paywalled and nobody publishes one — which is also why you should be wary of the regulation numbers quoted around this online. The ones I chased turned up only on software vendors' blogs and forum threads, and appear nowhere in the forms.

The certificate signed by someone who never came

There is a legitimate version of this, and most people don't know it exists.

An installer who isn't scheme-registered can appoint a third-party certifier before work begins, who inspects and tests the work and then issues a condition report — not an installation certificate. Three things make it lawful: the appointment happens up front, the certifier genuinely inspects, and what comes out is a report. The Welsh guidance is blunt about the boundary: an installation certificate can only be issued by the installer responsible for the work. (If you go looking for who currently offers this, be careful — the GOV.UK page listing third-party schemes was last updated in 2014 and still names a scheme that closed in 2021.)

Then there's the other version, and it does get prosecuted.

In October 2024 a Swindon trader was sentenced after Wiltshire Council's prosecution for, among other things, fraudulently issuing householders with an electrical installation certificate claiming the work had been signed off by an electrician who had not visited the property. He got a 12-month community order, 200 hours of unpaid work and £4,669.40 in compensation. A year earlier, Cheshire West and Chester prosecuted a man who advertised himself as NICEIC qualified when he never had been. And in March 2026 a forger was jailed for three years over a decade-long trade in fake qualifications, electrical installation among them.

The law behind those first two changed in April 2025: the Consumer Protection from Unfair Trading Regulations were revoked and replaced by the Digital Markets, Competition and Consumers Act, which lists displaying a trust mark without authorisation, and falsely claiming approval by a body, among practices unfair in all circumstances. One detail matters for anyone tempted to sign a favour for a mate: where the offence is down to another person's act, that person commits it too — whether or not they are a trader.

If you've been handed one and you're not sure

The most useful thing here is an admission from NICEIC itself: it cannot confirm whether a certificate is valid, because that would mean physically checking the work and cross-referencing the results.

What you can check is narrower and still worth doing. Whether the business is registered, through the competent person register — England and Wales only, which is a gap for Scottish readers — and whether the work was actually notified, through the schemes' notification lookup — though a record much older than six years may have dropped off it. Scotland has its own certification register.

The scheme guarantees are worth understanding before you rely on them. NICEIC's and NAPIT's both cap at £25,000, both run six years, and both are scoped to work done by a member within the scope of their registration at the time. Neither is a compensation scheme — they are about getting defective work put right. Which also means the guarantee is least likely to help in exactly the case you'd most want it to: work certified by someone who wasn't registered for it.

Where grafter.ly fits

Nowhere near the signature, and that's deliberate. We don't make cert software: no EIC comes out of grafter.ly, and your test results stay in whichever cert app you already run.

What it holds is everything either side of them — who asked for the work, what you quoted, when you were on site, what you billed. When somebody queries a certificate two years on, that trail is usually what settles it, because it shows who instructed what, and when. Writing the cert number onto the bill is the small habit that ties the two together, and our invoice template for electricians leaves a line for it.

---

Sources. The forms and the standard: the IET's model forms for BS 7671:2018+A4:2026, in particular the EIC and the EICR; IET Wiring Matters on competence and on EICR myths; the Electrotechnical Assessment Specification, June 2026; and Electrical Safety First's best-practice guides on condition reporting and test instruments. The law: the Building Regulations 2010, regulation 20 and Schedule 3; Approved Document P and the Welsh Government's Approved Document P; the Electricity at Work Regulations 1989, regulation 16, with HSE's guidance on those regulations and GS38; the Housing (Scotland) Act 2006, section 19B; the electrical safety standards regulations for rented homes in England and the Scottish Government's repairing standard guidance; Schedule 20 to the Digital Markets, Competition and Consumers Act 2024; assimilated eIDAS Article 25 on electronic signatures; and UKAS LAB 5 on how calibration due dates are reported. Schemes and checks: NICEIC's help hub and Approved Contractor scheme rules; GOV.UK's third-party certification page, last updated in 2014; the competent person register and the notification check. Enforcement: Wiltshire Council's report of the October 2024 sentencing (its page refuses automated access, so it isn't linked here) and Electrical Safety First's account of the same case; Cheshire West and Chester's 2023 prosecution; and the March 2026 sentencing in York. All checked 23 September 2026.

This is general information about how certification works, not advice on a particular certificate or a particular dispute.